Systems for processing documents are known. Such systems can receive electronic documents, e.g., as print data, or as physical documents. In operation, the system performs one or more processing steps on the received documents, e.g., in order to print the documents, assemble the documents into mail pieces, to provide the documents with a franking value or otherwise.
During the processing, the received documents are searched to find a barcode marking provided to the document. The barcode can, for example, be used for tracking and tracing purposes. Depending on the system, the settings of the search process are either set manually or automatically.
For example, European patent publication 1 347 372, the contents of the entirety of which are hereby incorporated by this reference, discloses a method and device for printing and preparing mail from a data stream in which the settings are set manually. From amongst the data stream an identification criterion for various pages is selected and the data stream is analyzed to identify the pages concerned. The identification criterion is selected by displaying one or more pages on a display, and selecting by the user a geometric position of the data in the data stream, which serves as an identification criterion. However, a disadvantage of the method and device known from European patent publication 1 347 372 is the manual selection by the user. This disadvantage is especially troublesome in case different types of documents are processed with the system, e.g., with different sizes or different locations of the marking. This requires a manual adjustment of the settings every time a new type of document is processed in order to ensure that the area provided with the document is scanned.
U.S. Pat. No. 4,034,341, the contents of the entirety of which are hereby incorporated by this reference, discloses an automatic postal-code number reading system. The system has a pre-scanning section and a main scanning unit. In operation, the pre-scanning section performs a coarse pre-scan on a relatively broad area of a mail item. The pre-scanning section then determines the location of a postal code within the relatively broad area and delivers x and y coordinates of this location to the main scanning unit. The main scanning unit scans the smaller area defined by the x and y coordinates with a fine scan. The data resulting from the fine scan are transmitted to a character recognition circuit in which the postal code is recognized.
However, a disadvantage of the system known from U.S. Pat. No. 4,034,341 is that the geometrical position of the broad area has to be pre-defined, i.e., the geometrical position has to be defined before the pre-scanning section performs the pre-scan. Furthermore, scanning the relatively broad area requires a relatively large amount of time.
United States Patent application 2005/0067496, the contents of the entirety of which are hereby incorporated by this reference, discloses a method and system for extracting graphical barcodes from template-based documents. A scanner reads a hard copy of a template based document to produce a digital scanned image of the hard copy. The image is processed by a graphical barcode extractor to determine the type of document template the hard copy is based on. The graphical barcode extractor determines a location of a barcode candidate from the determined type of document template. The image is then cropped based on the location of the barcode candidate, and the cropped image is decoded to extract information.
However, a disadvantage of this method and system is that the template has to be pre-defined before the processing of the image to determine the template. Furthermore, processing of the image to determine the template and a barcode candidate requires a relatively large amount of type and processing power.